This application outlines plans to develop artificial neural networks that will identify open reading frames in genomic sequence information. The investigators list eleven specific tasks that will train and design multilayer perceptron networks using an evolutionary optimization technique in a suitable user-friendly format. The investigators predict that these networks will provide a more effective computer-based system for automatic pattern recognition that will be useful in genome sequencing projects. PROPOSED COMMERCIAL APPLICATION: The commercial application will focus an development of software for gene recognition in unannotated DNA sequences.